Navigating Autonomy: Vision-Based Guidance for Two-Wheeled Inverted Pendulum Mobile Robots

In recent years, the convergence of Artificial Intelligence (AI) and Robotics has driven advancements in computer vision and autonomous systems, particularly within self-driving technology. The potential of autonomous robots to transform industries and societal norms is profound, yet their practical implementation hinges on safe and reliable navigation capabilities. This paper presents a novel approach to autonomous navigation for a two-wheeled inverted pendulum robot, leveraging custom computer vision models. By designing a navigation scheme tailored for a university campus environment, this study explores the integration of vision-based guidance with resource-constrained controllers to enhance robot autonomy. The paper addresses challenges in robot design, motor control, and course correction, proposing a solution based on Proportional-Integral-Derivative (PID) controllers and gyroscope feedback. Through experimentation and analysis, the effectiveness of the proposed navigation methodology is evaluated, highlighting the possibilities and limitations of vision-driven autonomy in robotics.